Ggtree notes

Setup

Libraries

library(tidyverse)
library(ggtree)
library(ggtreeExtra)
library(ape)
library(ggnewscale)
library(RColorBrewer)
library(svglite)

Paths

metadata_path <- 
    "data/processed/metadata_ashton_desj_all_weavepop_H99.csv"
duplications_path <- 
    "results/tables/duplications.tsv"
merged_tree_path <- 
    "data/processed/tree_merged.newick"

Include color palettes from R script

source("scripts/metadata_colors.R")

Example of color palette objects

lineage_colors

     VNBI     VNBII       VNI      VNII 
"#1B9E77" "#D95F02" "#7570B3" "#E7298A" 

Tree

Read tree

tree_original <- read.tree(merged_tree_path)

Metadata

Read and organize metadata.

To add metadata to a ggtree object (ggtree(tree) %<+% metadata), the first column of the metadata table must contain the labels in tree$tip.label.

metadata <- read.csv(
    metadata_path,
    header = TRUE) %>%
    select(strain, sample, source, lineage, country_of_origin, dataset)
head(metadata)

strain	sample	source	lineage	country_of_origin	dataset
20427_2#42	ERS1142739	Clinical	VNI	Vietnam	Ashton
04CN-64-065	ERS2540972	Clinical	VNI	Uganda	Ashton
BMD915	ERS2541156	Clinical	VNI	Vietnam	Ashton
14892_1#7	ERS542302	Clinical	VNI	Vietnam	Ashton
BMD942	ERS2541105	Clinical	VNI	Vietnam	Ashton
BMD2801	ERS2541198	Clinical	VNI	Vietnam	Ashton

Manipulate tree

Remove tips that are not in metadata$strain

tree <- drop.tip(tree_original, setdiff(tree_original$tip.label, metadata$strain))

Color tips

Color the points of the tips with geom_tippoint() This works even when not all tips are in the metadata table, NA values are used when needed.

ggtree(tree_original, layout = "circular") %<+%  metadata +
    geom_tippoint(aes(color = lineage), size = 2)+
    scale_color_manual(name="Lineage", values= lineage_colors)

Color labels

Add the tip labels (strain) and color them by a column in the metadata.

ggtree(tree, layout = "circular") %<+%  metadata +
    geom_tiplab(aes(color = source), size = 2, align = TRUE)+
    scale_color_manual(name="Source", values= source_colors)

Identify basal nodes of clades

To identify the most basal node of each lineage, get the Most Recent Common Ancestor of a pair of strains in the lineage, they have to be two strains whose MRCA is at the base of the clade. You can also use this to identify any clade of interest.

VNI_node <- getMRCA(tree, c("Tu241-1","UI_31647-2"))
VNII_node <- getMRCA(tree, c("C2","C12"))
VNBI_node <- getMRCA(tree, c("Tu229-1","Ftc267-2"))
VNBII_node <- getMRCA(tree, c("MW-RSA3321","MW-RSA3179"))

nodes_lineages <- data.frame(
    lineage = c("VNI", "VNII", "VNBI", "VNBII"),
    mrca = c(VNI_node, VNII_node, VNBI_node, VNBII_node)
)

Get the node number of the MRCA of each VNI sublineage

VNIa4_node <- getMRCA(tree, c("04CN-30-008","UI_31647-2"))
VNIa5_node <- getMRCA(tree, c("BMD852","04CN-64-014"))
VNIa93_node <- getMRCA(tree, c("04CN-65-080","04CN-65-002"))
VNIa32_node <- getMRCA(tree, c("BMD942","BMD2801"))
VNIaX_node <- getMRCA(tree, c("Bt48","04CN-63-007"))
VNIaY_node <- getMRCA(tree, c("04CN-65-073","Bt138"))
VNIb_node <- getMRCA(tree, c("04CN-65-096","MW-RSA722"))
VNIc_node <- getMRCA(tree, c("Bt20","Bt11"))

nodes_sublineages <- data.frame(
    sublineage = c(
                "VNIa-4", "VNIa-5", "VNIa-93",
                "VNIa-32", "VNIa-X", "VNIa-Y",
                "VNIb", "VNIc"),
    mrca = c(
            VNIa4_node, VNIa5_node, VNIa93_node,
            VNIa32_node, VNIaX_node, VNIaY_node,
            VNIb_node, VNIc_node))

To print the tree with all the node labels:

ggtree(tree, layout = "circular")+
    geom_text(aes(label = node), size = 3, vjust = -0.5)

Add clade labels

ggtree(tree, layout = "circular") +
    geom_cladelab(data = nodes_lineages, 
                mapping = aes(node = mrca, label = lineage, color = lineage))+
    scale_color_manual(name="Lineage", values= lineage_colors)

Add node labels

ggtree(tree, layout = "circular") +
        geom_text(aes(label = nodes_lineages$lineage[match(node, nodes_lineages$mrca)]), 
                        size = 4, , fontface = "bold",
                        hjust = 1.25, vjust = -0.5)

Color branches

To color the branches of clades, create groups in the tree object with groupClade().

tree <- groupClade(tree, setNames(nodes_sublineages$mrca, nodes_sublineages$sublineage))

And color the tree by aes(color = group). Use scale_color_manual() with a palette to specify the colors of the clades and avoid coloring the branches that are not part of the selected clades.

ggtree(tree, aes(color = group), layout = "circular", 
    size = 1) +
  scale_color_manual(
    name = "Sublineage",
    values = sublineage_colors
  )

Color branches and tips and use node labels

ggtree(tree, aes(color = group), 
        layout = "circular", size = 1) %<+%  metadata +
    scale_color_manual(
        name = "Sublineage",
        values = sublineage_colors)+
    new_scale_color()+
    geom_text(aes(label = nodes_lineages$lineage[match(node, nodes_lineages$mrca)]), 
                        size = 4, , fontface = "bold",
                        hjust = 1.25, vjust = -0.5)+
    geom_tippoint(aes(color = source), size = 1)+
    scale_color_manual(name = "Source", values = source_colors)

ggtree(tree, layout = "circular") %<+%  metadata +
    geom_cladelab(data = nodes_sublineages, 
                mapping = aes(node = mrca, label = sublineage, color = sublineage),
                align = TRUE, face = "bold")+
    scale_color_manual(name="Sublineage", values= sublineage_colors)+
    geom_text(aes(label = nodes_lineages$lineage[match(node, nodes_lineages$mrca)]), 
                        size = 6, , fontface = "bold",
                        hjust = 1.25, vjust = -0.5)+
    new_scale_color() +
    geom_tippoint(aes(color = source), size = 2)+
    scale_color_manual(name = "Source", values = source_colors)

ggtree(tree, layout = "circular", size = 1, branch.length = "none") %<+%  metadata +
    geom_hilight(data=nodes_sublineages, aes(node=mrca, fill=sublineage), alpha = 0.5)+
        scale_fill_brewer(name = "Lineage", palette = "Dark2")+
        new_scale_fill()+
    geom_cladelab(data = nodes_sublineages, 
        mapping = aes(node = mrca, label = sublineage),
        align = TRUE, face = "bold",
        angle = "auto")

Add data as heatmap

Create a dataframe where the rownames are the tree’s tip labels, each column is a column of the heatmap and the values are the colors in the heatmap. If using categorical values for the heatmap and their order in the legend is important, convert the column to factor and set the order of the levels.

country <- metadata %>%
    select(strain, country_of_origin)%>%
    column_to_rownames("strain")

country$country_of_origin <- factor(country$country_of_origin,
                                levels = names(country_colors))

To add a heamap, the function gheatmap() takes an object made with ggtree() and add the dataframe with the matrix to use in the heatmap. Use the offset option to separate the heatmap from the tree. Use the guides() function to order the legends. Use the option limits in scale_fill_manual to set the order of the legend labels to match the order of the levels in the factor.

p <- ggtree(tree, branch.length = "none")  %<+%  metadata
gheatmap(p, country, width=0.1, colnames = FALSE,  
            colnames_position = "top", 
            offset=7,
            font.size = 6) +
    geom_tiplab(aes(color = lineage), size = 3)+
    scale_color_manual(name = "Lineage", values = lineage_colors)+
    scale_fill_manual(values=country_colors, 
                      name="Country", na.translate=FALSE,
                      limits = levels(country$country_of_origin))+
    guides(color = guide_legend(order = 1), fill = guide_legend(order = 2))+
    theme(legend.position = "right")